Someone told me last week that AI was going to replace what I do. I’ve spent twenty years and 250+ annual reports turning board papers, interviews and half-formed ideas into something people can actually understand. Apparently “prompt engineering” makes all of that obsolete.
AI Is Literal. Your Instructions Need to Be Too
AI is literal. It processes patterns; it can’t read your mind (yet). The quality of what it gives back depends entirely on what goes in, and what goes in is a prompt: specific, clear, structured, unambiguous.
That’s not a technical skill. It’s the same discipline plain English writing has demanded for decades — say exactly what you mean, in the fewest words that will do the job.
Poor instruction: It would be beneficial if some consideration could potentially be given to reviewing the existing documentation.
Plain English: Please review the existing documentation.
Poor prompt: Write something about our annual report.
Plain English prompt: Write a 150-word summary of our FY26 annual report for a first-time investor, focusing on revenue growth and our sustainability commitments.
AI responds exactly the way a rushed, overworked colleague would, the clearer the brief, the better the output. Vague instructions get answered with blistering confidence anyway which is often worse than getting no answer at all.
Active Language and Structure Still Win
Plain English writing has always demanded active voice and short, direct sentences. The best AI prompts use exactly the same structure, for exactly the same reason: passive voice leaves too much room for interpretation.
Instead of: The campaign was monitored by the project team.
Write: The project team monitored the campaign.
In annual report writing, this isn’t a style preference — it’s what makes a report’s key messages land instead of getting lost. The same logic applies whether you’re briefing a writer or briefing an AI model.
Context Is Not Optional
A good writer gives the reader enough context to understand the situation without having to guess. A good prompt does the same.
Instead of: Write an article about marketing.
Write: Write a 1,000-word article for small business owners explaining how SEO and Google Ads work together to generate leads.
Structure produces better results too. Rather than asking for a proposal that “explains the costs and benefits and includes examples and is persuasive,” a structured brief looks like this:
Write a proposal that:
- Explains the problem
- Outlines the solution
- Includes three examples
- Provides a cost-benefit analysis
- Ends with a recommendation
This is how effective business writing has always worked. Humans and AI both make assumptions when instructions are vague — the difference is that AI won’t ask a follow-up question. It will simply guess, and answer as if it were certain.
Structure Produces Better Results
Professional communicators break information into logical sections. Rather than: “Write a proposal and explain the costs and benefits and include examples and make it persuasive”, a structured prompt would be:
Write a proposal that:
- Explains the problem
- Outlines the solution
- Includes three examples
- Provides a cost-benefit analysis
- Ends with a recommendation
This mirrors how effective business writing has always worked.
Humans and AI both make assumptions when instructions are vague.
AI Prompt Engineers Are Plain English Masters
Strong prompt engineering isn’t a new technical discipline. It draws on skills professional communicators have always needed:
- Clear writing
- Logical thinking
- Process documentation
- Communication
- Brief writing
- Technical writing
- Instruction writing
The professionals getting the best results from AI are, almost without exception, the same people who were already good at plain English writing — distilling complex ideas down to their essential parts. Writing talent and prompting talent are the same talent. The reports of the death of good writers have been greatly exaggerated. If anything, AI has only highlighted their value.
We Didn’t Just Write About This, We Built a Product On It
After 250+ annual reports and other complex publications, tight, compelling content is simply what we do, whether we’re crafting a report or a watertight AI prompt. Writing well is what lets you use AI properly. It’s the same skill either way.
That thinking is built directly into ReportNavigatorAI™, our proprietary platform, made the same way we make our reports: precise, verifiable, worth trusting.
ReportNavigatorAI™ sits on top of your published annual or sustainability report and lets readers interact with it using natural language. Users can:
- ask questions in plain English
- retrieve precise answers drawn only from the report
- see exact page citations for every response
- jump directly to the source page
- explore report data through interactive dashboards
- speak or type questions in 50+ languages
Every answer is linked back to the original report page for verification. No hallucinations. No approximation. Just verified information from your published report.
It’s the same discipline in reverse: instead of demanding clarity from the writer, it brings clarity to the reader.
A Report That Earns Its Keep All Year
Instead of a static PDF gathering digital dust, your report can now work all year round — externally for investors and stakeholders, and internally as a searchable knowledge base. Need a figure from last year’s report? A few keystrokes, not a busy EA flicking through a PDF.
It’s early days, but our clients already range from federal government departments to housing providers to ASX- and NSX-listed companies.
Use AI to Improve Your Next Annual Report
None of this starts with a clever tool. It starts where good communication always has — with someone who knows how to say precisely what they mean.
At Andrew Pegler Media, we bring that discipline to your annual report writing and to the AI tools we build, including ReportNavigatorAI™.
To find out more, contact us or call 1300 209 849.
